Bosses Want You to Use AI but They’re Not Setting a Good Example, Study Says

Your company’s leadership might’ve told you to start incorporating AI into your work, but that same leadership may actually be hindering AI adoption across the workplace, according to a new report from Microsoft.

Microsoft’s Work Trend Index, published Tuesday, tracks changes in workplace tech, behaviors and culture. This year’s report, based on global survey data and real-world data from Microsoft customers, focuses on how companies are incorporating AI. Many AI users (65%) say they fear falling behind if they don’t adopt AI quickly, but 45% say it feels safer to stick to current goals than to redesign their workflows. Very few (13%) feel rewarded for their AI innovation.

This report highlights a new facet of the debate over how AI could be used in the workplace. For years, executives have been pushing their employees to integrate AI so they can say their companies are on the cutting edge — even in cases where AI hasn’t been proven to be useful or has worsened employees’ work-life balance. Companies have been implementing layoffs under the pretext of replacing employees with AI while also pushing staff to beef up their AI literacy and skills. 

Now, Microsoft is reporting seeing a “bottoms-up groundswell in AI fluency,” Matt Firestone, general manager of product marketing for Copilot, told me.

Microsoft analyzed more than 100,000 de-identified chats with Copilot and found nearly half (49%) involved employees asking for help with “cognitive work” — tasks like analyzing information, solving problems and thinking creatively.

The number of AI agents in use has grown 15 times year over year. AI agents are customizable bots that can handle tasks independently. They’re largely seen as the next wave of generative AI and use the most advanced AI models.

We’ve seen AI disruptions across the board, from legacy tech companies to entertainment giants. But promises from leaders like Nvidia CEO Jensen Huang of smaller human workforces overseeing armies of millions of AI agents haven’t yet come to pass. One possible explanation, Firestone said, was based on an old adage: Culture eats strategy for breakfast.

“If you can change processes and culture to unlock [employees’] potential, our belief is that’s how technology will diffuse through an organization a lot quicker,” Firestone said.

Leaders need to set a better AI example

Only 26% of AI users surveyed say their leadership is clearly and consistently aligned on AI. Others report limited capacity or agency — employees may not have the tools or programs they need to implement AI, or they have the skills but can’t use them. A lack of organizational support can also mean that employees who are told to start using AI don’t know quite where to begin.

One of the biggest recommendations in the report is for managers to model effective AI use, showing employees which uses are acceptable and actually helpful. In a 2025 Microsoft survey, managers who modeled AI use led to a 30-point increase in employees’ trust in agentic AI.

“It’s this human instinct,” Firestone said. “If I see someone doing [using AI] and sometimes being successful and not being successful, that experimentation makes me more comfortable about being in the open about it.”

It’s one thing for managers and executives to issue mandates to use more AI — it’s entirely another challenge for employees to find useful ways to do so. Microsoft’s report highlights that some employees want to dive into incorporating AI and agentic AI into their work, but there isn’t the necessary support or resources to do so effectively. That is ultimately a leadership problem, not just a technical one.